Rules of the genetic code are established by the matching of trinucleotide sequences with specific amino acids. This is done by the interaction of aminoacyl tRNA synthetases with transfer RNAs. This specificity of this interaction enables the synthetases to attach amino acids to the cognate transfer RNAs. The transfer RNAs encode nucleotide triplet anticodons which correspond to specific amino acids. For some, perhaps most, transfer RNAs, the anticodon per se is not the primary site by which the synthetases recognize their cognate transfer RNAs. This means that other sites in the molecule are critical. Mutagenesis of tRNA and its cognate enzyme is used to define further the sites that are sensitive to recognition. Advantage is taken of structural data on transfer RNA and of emerging structural data on aminoacyl tRNA synthetases. Also used are data which have defined regions in both molecules that are significant for recognition. Chimeric enzymes are tested to determine whether specific regions of one enzyme can function within the context of the structural framework of another enzyme. Sequences that are critical for tRNA recognition may be swapped between enzymes. The use of heterologous yeast-E. coli systems in vivo expands considerably the ability to define nucleotides that determine the identity of a tRNA. These systems also can be developed to make chimeric yeast-E. coli aminoacyl tRNA synthetases.